Giving the single clade timeseries workflow a community with multiple clades causes the number of clades to become the analysis' number of samples.
This error can be reproduced by using a 3 clade community and 4 time points. (num_samples != number of clades) The workflow progresses up to deconvolution, where the numbers get mixed up (so to speak). Downstream at measure_accuracy.py the incorrect number of strains is used and you can get indexing errors.
Either an assertion is overlooked at a Process level or in munging channels.
Giving the single clade timeseries workflow a community with multiple clades causes the number of clades to become the analysis' number of samples.
This error can be reproduced by using a 3 clade community and 4 time points. (num_samples != number of clades) The workflow progresses up to deconvolution, where the numbers get mixed up (so to speak). Downstream at measure_accuracy.py the incorrect number of strains is used and you can get indexing errors.
Either an assertion is overlooked at a Process level or in munging channels.